27 replies to this topic

[Kanatta Jing]

kesmai, on 25 November 2014 - 01:22 AM, said:

easy. If you are winning 20 times in a row: Elo toO low.

if you are losing 20 times in a row: Elo too high.


o7

If you lost 20 games in a row it probably isn't because you're ELO is to high. The following are more likely examples of conditions that could lead to such a phenomenon:

tired
hungry
itchy
cranky
drunk
viscously emasculated by your loved ones
thirsty
migraine
sick
withdrawal

If you are winning 20 games in a row it is likely because of the sudden removal of one or more of the above conditions.

[kesmai]

Sounds also reasonable. But you forgot to take the moonphase into calculation.

[LORD ORION]

Did you fire 2000 LRMs and do 100 dmg during the whole match?
Your elo is high now, stop using LRMs.

[ice trey]

Colin Thrase, on 24 November 2014 - 09:52 PM, said:

with CW on its way

If it's something playable by December 16th, I can get a King Crab. If it's not playable by then, I go the C-bills route to get my Panther.

I have a strong feeling it's the latter.

[Karl Streiger]

So - how to estimate ELO:
first you have to collect all games for each weight class.
for example Medium: 500 Wins, 400 Losses.
67% of the wins are a result of the MM (it determine that you may win) - the other 33% was a win against the odds.
Those 165 will result in 21-40 points of additional ELO (based on 50-250 ELO difference and its unlikely that you have won a match with 250 difference rather than luck)
So 165 games rise the ELO

Now the looses - same here 67% may the looses where you were supposed to loose because your side hat the lower ELO.
The other 33% you did loose although you should have won - will result in a loose of 21-40 - lets say 20 on the average.
132 games reduced the ELO

Now the value of rising and lowering ELO results in the differences between the teams. Russ stated that they had a different ELO of 50 in solo queue and up to 250 in group queue - but that means there are games that had much lower differences and bigger difference.
A win against a much better ELO team result in a major boost of ELO, while the loose vs a almost equal team will hardly change your ELO.

If you say that on the average the ELO difference was 50 - your ELO may have risen by 165-132 = 33 * 20 Points = 660.
So your ELO must be near 2000 - so you are in the upper part of the average - but you should already know that by the W/L ratio of 1.25 W/L ratio.

Thats one theory.
The other: even if you have a W/L of 0.5 doesn't mean you ELO is near the newbs and noobs. maybe you did loose only those games you were supposed to loose - and you beat great teams during you win - even when the MM placed you on the "looser" team.

Well - of course there are the "other" players. and that is the problem - if you think / know that your ELO can't be superb - and you see a group of "maybe" top players in the other team:
Two things may have happen: You have top players in your team too, the other players in the other team have a low ELO.

Edited by Karl Streiger, 25 November 2014 - 03:16 AM.

[Fiona Marshe]

When the MM consistently throws you a 300 ton disadvantage, it thinks your ELO is too high (even when the W/L drops below 40%)

[Water Bear]

This sounds like a bad idea to me. Elo rating is probably a lot like IQ -- there isn't much you can do to improve it but it's easy to feel bad about.

Also be aware that no one seems to have any idea how it works. I haven't read all the replies in this thread yet but the very first one made claims I don't think the poster could verify.

Edit: Oh, also. There's a wikipedia article about the Elo rating system. The basic operation is very simple -- if you beat someone with a higher Elo rating, you gain points and they lose points. After that I know nothing about how the Elo system is (or is not) implemented in MWO. (And probably neither does anyone else, it seems...)

Edited by Water Bear, 25 November 2014 - 06:19 AM.

[Angel of Annihilation]

Vassago Rain, on 24 November 2014 - 10:10 PM, said:

Sure.

Do you go up against known community qualities, like LORDS, steel jags, and etc?
Do you find that most of your matches are ggclose clean sweeps (an indication that elo is actually working)?
Do you have to carry hard, or carry softly?
Do you think that LRMs are a competitive weapon?
Do you know how the game's mechanics work?

Things like that.

Things like that are worthless. Why?

Because the game changes at different ELO brackets. First we really have no clue how good Lords or Steel Jags are because we don't get to see their ELO. People just assume they are the elite players in the game but they may very well only be elite for the ELO bracket they play in.

Also if you are generally at the top of your ELO, even if you are a poor player overall, your still going to end up carrying matches just because everyone else your fighting with and against is a poorer player than you.

Lastly, weapons and play styles requirements change at different ELOs. For example, whatever ELO I play in was never friendly to jump snipers and poptarts. Generally they got shot out of the sky very quickly or at the very least suppressed. Now whether that means they weren't skilled enough to jump snipe correctly or that we were so skilled at targeting and firing at these guys as they made themselves a big fat target in the air, I don't know, but Jump Sniping was never META in the matches I played.